greetings,
pardon my ignorance, yet...
hasn't someone come up with a way to "cycle in" or "soft reboot" a multiprocessor linux machine without losing *uptime*... um the same way you would in a mission critical telco environment or whatever? :)
- rh